Coronavirus in Context: Using Data to Address COVID-19 Challenges

[MUSIC PLAYING]

As we begin to reopen, we need to have good data. I sat down a little while ago with Murray Aitken, senior vice president of IQVIA, and Bob Brisco, CEO of Internet Brands and WebMD, to talk about the type of information we need.

Could we have some type of reopening index? And what's the role of human data science as we look at trends and try to make decisions?

Thank you both for joining me. Murray, you have this concept of a reopening index, where you've talked a little bit about moving somewhat from a national perspective to a sub-national, looking at, well, what are the dynamics in the local community? Can you explain that a little more?

Sure. And thanks for the opportunity to talk about our favorite topic here, which is really how to, uh, grab the, uh, the-- the data science capabilities that we have, integrated with what we know about human science-- we call that human data science-- and apply that. We quickly need to get beyond talking at national levels and talking at sub-national levels.

There is a question as to whether the sub-national level should be states or even within states. Or we have another way of drawing a map of the United States, which is really based on what we observe every year with flu season, where we see the pockets of activity that transcend state lines. And we actually think this is a more useful way to talk about sub-national units.

That being said, not much gets measured on that basis. But we think it's important, as different parts of the country re-- "reopen," quote, unquote, in different ways and at different paces, that we're able to look at the impact that-- that-- that-- that that is having. So hence, our-- our call for some type of reopening index. And I can talk about what I think the elements could be in that. And that we should be looking at that at a sub-national level.

Murray, can you tell us more about this human data science? That's not a term that most people are familiar with.

Right. Well, thanks for that question, because it's my favorite topic. Uh, human data science is a combination of three familiar words that really haven't been put together in-- in the way-- in this way in the past. And what-- the way we think about it is the intersection of human science and data science.

So if I can start with, uh, data science, we see that there's a lot going on. Everyone's talking about AI and big data and-- and, you know, applying algorithms, and so on. You know, first off, that's not straightforward. And it's definitely not straightforward in-- in health care. Um, accessing reliable and clean and linkable data is, as we say, not for amateurs. And-- and so there's an enormous amount of, um, uh, activity and focus in order to really elevate, uh, the-- the capabilities and role that data science can play.

But then you've got to link that to human science. What do we really understand about human biology, about disease ideology, um, about the progression of disease, about the impact that interventions, um, have on-- on the human body, and outcomes, and so on. Uh, you know, there's still a lot of drugs that get dispensed every year where we don't really understand exactly how they work. We just sort of know that they do. Right?

So there's-- there's still enormous gaps, uh, from a human science perspective. Data science together can actually help enhance that understanding. And it's that linkage that we think is very valuable.

There's one other angle. And that's about human data. So we make a point of talking about human data as opposed to-- as-- as supplemental, in a sense, to patient data. So patient data, we define as information that's gathered when an individual interacts with a health system. And we know that that happens periodically. Um, and-- and it's-- it's an incomplete picture of that individual as a human. We also know that outcomes are very much driven not only by the quality of services that a patient receives or the drugs that they may be prescribed, but also a lot of other factors-- social determinants, behavioral factors, and so on.

Uh, we believe that, in order to really advance our understanding, um, of health care, we need to be pulling in a lot more of that human data, um, into the data science and human aspects into understanding human science. So that, sort of in a nutshell, is what we're all trying to bring together in talking about human data science.

In-- in your view, is there a gap that's fairly significant now?

Yes. There's a huge gap in, um, A, the, uh, quality and extent of the data that is being accessed and worked on that's coming out of the health system. There's still large gaps in the data. There's large interoperability issues, as-- as you probably well-- well recognize. And there's huge gaps in terms of not capturing, um, the non-clinical data, the-- the non patient interacting with the health system information. Right?

We-- our EHRs, generally, are not capturing anything about the lifestyle of the patient or their social conditions or their work or home environment or their food supply, and so on. Yet, we know that those have an enormous impact on outcomes across a whole range of, um, diseases. So I think those are-- those are big gaps that, um, we really do need to close if we're really going to, you know, make the progress we need.

And-- and as those gaps get filled in, do you think it unleashes a lot of insight that will--

Yes.

--save lives?

Yeah, absolutely. Um, you know, that being said, um, as I also remind our data scientists, uh, the value of data science, uh, no matter how great the data, is only def-- only, uh, occurs-- the value is only delivered when someone makes a different decision or takes a different action.

And another gap that I feel is-- we've got a lot of great algorithms that are yielding great insights. But does that actually change behavior and change, uh, uh, uh, decisions, and therefore, change outcomes? And I think that's actually a growing, uh, gap, because there's not enough, uh, emphasis on, what does it really take, uh, to make that kind of a change? And it's more than just spitting out, you know, the answer to, uh, to-- to an algorithm or-- or a sort of quantitative output.

Is-- is-- Murray, is another leg of this stool the-- I-- I don't want to go too far with this-- but the-- the patient's data themselves. And I'm not going into the extreme quantified self, but self-reported health loops--

Yeah.

BOB BRISCO: --and post-care loops, and, of course, um, devices that get other metrics would be part of that. But it seems like-- is there-- is there a lot of opportunity--

Yes.

--in that area?

Yes. And I think the evidence for that is growing. And by the way, COVID-19 is bringing more of that evidence to the fore, because we've got more people, um, you know, self-monitoring, self-reporting, remote monitoring, therefore giving us more data points, um, about the, uh, individual. And as those data points, uh, expand, so do the insights.

For example, what is the optimal time for someone to come back following their, uh, cancer treatment for a checkup? Um, you know, that timing is not really evidence-based. And we reduce the-- the sort of-- if we can reduce the barrier of someone coming back for the checkup, and make it more self-reported or, you know, that there's more continuous monitoring, then I have a feeling that's also going to change the frequency of checkups, optimize it more, and result in better outcomes. And again, there's-- there's a growing amount of evidence that that is the case.

Yeah. We're seeing in our, um, Krames business that we recently acquired and has discharge instructions for thousands of hospitals, as we tighten up that link, there are gems in there about the self-reporting, that if the clinicians understood them, the appropriate intervention at the appropriate moment can have-- be efficient and relatively inexpensive to do and power-- and have powerful effects. But--

Yes.

--the gaps in knowledge there are fairly big. But they also seem relatively easy to close with digital technology now.

Right. That's right. I'm-- I'm reminded, though, of the, uh, anecdote I heard when-- when all the focus was on reducing readmissions to hospitals. And somebody made the observation that, you know, for all-- again, for all the sort of efforts to, uh, you know, tackle that issue, the person who could best predict, uh, readmission was the discharge nurse, who was the last person who saw the patient, saw who, if anyone, came to collect the patient, saw the means of transportation, and, uh, from that information, in a human way--

Yes.

--was actually the best equipped--

And-- and--

--to predict re-- readmission.

And that brings in the whole concept of-- of the care loop and the care community--

Yeah. Absolutely.

--and all of that.

Yeah. Absolutely. Yeah.

[MUSIC PLAYING]